Authorship After Origin: Accountability, Relational Contracts, and AI Co-Production

This paper challenges the idea that authorship is defined by who produces a text. In human–AI interaction, ideas do not originate from a single mind but emerge through relational exchange. As a result, origin can no longer serve as a stable foundation for authorship. Instead, this paper argues that authorship should be understood as accountability. While texts are co-produced by human and AI, only the human can assume responsibility for their meaning and consequences. This asymmetry does not weaken authorship—it clarifies it. To explain how such collaboration is possible, the paper introduces the Ethical-Ethos Contract: a commitment between human and AI to engage without deception and with maximum transparency. Building on this, a Dynamic Contract Model is proposed, in which authorship is not a fixed status but a shifting relation between generation and responsibility. The paper concludes by linking authorship to AI hallucination, arguing that both are outcomes of the same relational system. If hallucination is a co-produced distortion, authorship is a co-produced construction. The question is no longer who created the text, but who stands behind it.